VAE-Based Rhythm Disturbance Index Correlates with Bilateral Symmetry Breakdown in Human Motion

Rhythm disturbances during human exercise represent a critical challenge for both physiological monitoring and athlete safety. To address this, a structure-enhanced β-TCVAE framework was proposed that derives a Rhythm Disturbance Index (RDI) from multimodal wearable sensor signals. RDI demonstrated a strong correlation with bilateral imbalance (r = 0.838, R2 = 0.702) and achieved high discriminative performance (ROC-AUC = 0.823). Importantly, its weak and non-significant correlation with heart rate (r = 0.0569, p > 0.05) supported independence from cardiovascular load, underscoring its specificity to motor rhythm rather than systemic exertion. Analyses conducted on multimodal datasets further validated the robustness of this correlation, showing that RDI consistently aligns with disruptions in locomotor symmetry even after controlling for heart rate. This quantifiable coupling between rhythmic instability and symmetry loss positions RDI as a dual correlational indicator, sensitively reflecting both neuromuscular rhythm irregularities and axial imbalance. Such dual insight enables continuous and objective monitoring of locomotor quality, empowering coaches, clinicians, and sports scientists to tailor training strategies, optimize performance, and reduce the risk of injury. By integrating advanced variational reasoning with real-time wearable sensing, the proposed framework offers an evidence-based step forward in precision monitoring and risk assessment for athletes.